UP2DATE completes its first milestone

UP2DATE completes its first milestone

After 6 months from its kick-off, the UP2DATE project has completed its first milestone and has successfully accomplished its goals. The efforts in this initial phase have focused on the definition of project requirements and success criteria, the selection and setup of the hardware and software platforms and on an initial safety and security analysis of such platforms.

Progress at MS1

The definition of the requirements was based on the objectives of the project as well as on the individual needs of the project use cases, which had a crucial role in this initial project phase. As a result, the consortium has defined the high-level building blocks of the railway and automotive use-case architectures, a research scenario, and an extensive list of requirements.

Together with the requirements, the selection of the hardware platforms as well as their software stack and related tools are a crucial result of this first milestone. To this end, the consortium has explored different heterogeneous high-performance platforms that are capable of functional safety certification. The decision has been taken based on the satisfaction of several requirements such as, performance, portability, observability, partitioning, or requirements from end-users. In short, the project will focus on a variety of 4 platforms: 2 decided by direct requirements of the industrial users (Renesas RCar H3 and Infineon Tricore TC397) and 2 research heterogeneous platforms (the NVIDIA Xavier and the Zynq Ultrascale+).

The latter has been decided through careful inspection of the candidate platform manuals as well as with extensive experimentation and benchmarking. On the one hand, the Xilinx Zynq Ultrascale+ is an attractive platform for the project due to its rich set of functional safety and security features that are complemented with extensive public documentation. On the other hand, the NVIDIA Jetson family offers at this moment the most powerful embedded general-purpose GPUs. The initial experimental results, which have been based on a Benchmark suite provided and adapted by the BSC, give an insight of the performance capabilities of each platform, where we have seen that the NVIDIA Xavier is the most powerful platform.